Educational data mining provides a way to predict student academic performance.

A Study of Educational Data Mining: Evidence from a Thai University

(https://doi.org/10.1609/aaai.v33i01.3301734)

This paper compared the student’s performance predic-
tion in two different aspects, access duration and access
frequency. On average, the analysis of internet access dura-
tion reveals a better accuracy than frequency analysis. It
means that if we know how long students spend on the
activities, this can provide a better accuracy.

By detecting students who are at-risk of failing their stud-
ies, this study provides insight for educators and govern-
ments to plan and reduce the cost of the university studies.
Analysing data from questionnaires by using statistical
techniques such as SEM (Vuttipittayamongkol 2016) are
useful and popular methods. However, the increasing
number of the volume and varieties of the datasets may
involve other techniques such machine learning techniques
to discover complex patterns. These techniques also offer
other benefits with reducing human intervention. Compar-
ing the accuracy of results between decision trees, Naïve
Bayes, Logistic Regression and Neural Network, the Ran-
dom Forest approach is found to be best suited for this type
of dataset.